Optimized Traffic Scheduling for AI Acceleration // Optimized Traffic Scheduling for AI Acceleration
ABG-132748
ADUM-60115 |
Thesis topic | |
2025-07-02 |
Avignon Université
AVIGNON - Provence-Alpes-Côte d'Azur - France
Optimized Traffic Scheduling for AI Acceleration // Optimized Traffic Scheduling for AI Acceleration
- Computer science
Traffic engineering, Reinforcement Learning , Scheduling ,
Traffic engineering, Reinforcement Learning , Scheduling
Traffic engineering, Reinforcement Learning , Scheduling
Topic description
Le Laboratoire d'Informatique de l'Université d'Avignon (LIA), Avignon, France, propose un poste de doctorant sur le thème « Optimized Traffic Scheduling for AI Acceleration » dans le cadre du projet ANR Net4AI (Network acceleration for Generative AI). La thèse se concentre sur le développement de modèles markoviens pour les charges de travail du trafic de l'IA. Les modèles reproduiront la récurrence des tâches d'IA en tenant compte de différents degrés d'information et de stratégies de basculement qui peuvent réaffecter les ressources au moment de l'exécution. L'objectif sera de définir des politiques appropriées pour rendre efficace l'utilisation de la communication collective, par exemple via une segmentation efficace des flux de traffic [Fdp18], ou en développant des politiques de routage qui prennent en compte le basculement. La présence de contraintes liées à l'infrastructure du réseau nécessitera l'utilisation d'outils sûrs et efficaces.
------------------------------------------------------------------------------------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------------------------------------------------------------------------------------
The Laboratory of Informatics of the University of Avignon (LIA), Avignon, France, offers a PhD position on “Optimized Traffic Scheduling for AI Acceleration” in the framework of the ANR Project Net4AI (Network acceleration for Generative AI). The thesis focuses on developing Markovian models for AI traffic workloads. The models will reproduce the recurrence of AI tasks accounting for different degrees of information and failover strategies that may reassign resources at runtime. The objective will be to define suitable policies to render efficient the use of collective communication, e.g., via efficient segmentation of traffic flows [Fdp18], or developing routing policies which are failover-aware. The presence of constraints related to the network infrastructure will require the usage of safe
------------------------------------------------------------------------------------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Début de la thèse : 01/10/2025
------------------------------------------------------------------------------------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------------------------------------------------------------------------------------
The Laboratory of Informatics of the University of Avignon (LIA), Avignon, France, offers a PhD position on “Optimized Traffic Scheduling for AI Acceleration” in the framework of the ANR Project Net4AI (Network acceleration for Generative AI). The thesis focuses on developing Markovian models for AI traffic workloads. The models will reproduce the recurrence of AI tasks accounting for different degrees of information and failover strategies that may reassign resources at runtime. The objective will be to define suitable policies to render efficient the use of collective communication, e.g., via efficient segmentation of traffic flows [Fdp18], or developing routing policies which are failover-aware. The presence of constraints related to the network infrastructure will require the usage of safe
------------------------------------------------------------------------------------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Début de la thèse : 01/10/2025
Funding category
Funding further details
Contrat doctoral
Presentation of host institution and host laboratory
Avignon Université
Institution awarding doctoral degree
Avignon Université
Graduate school
536 Agrosciences et Sciences
Candidate's profile
Français : Profil du candidat : le poste s'adresse à un candidat très motivé, titulaire d'un master en ingénierie des télécommunications, en mathématiques appliquées ou en physique.
Des compétences en communication écrite en anglais sont indispensables. Les candidats qui postulent prouveront et/ou justifieront les connaissances et compétences demandées en fournissant :
● Un curriculum vitae complet ;
● Une lettre de motivation cohérente avec le projet de doctorat proposé ;
● Tous les documents attestant des compétences et connaissances demandées ;
● Les dossiers académiques complets et les notes obtenues ;
● Une ou deux lettres de recommandation sont considérées comme un plus.
Compétences requises : Le cursus du candidat doit démontrer de solides compétences théoriques en modélisation et en évaluation des performances. La connaissance de la théorie du contrôle, de la théorie de Markov et de l'apprentissage automatique (en particulier l'apprentissage par renforcement, les réseaux neuronaux et l'apprentissage fédéré) est préférable. Une expérience en matière de simulations événementielles est considérée comme un atout.
English: Candidate profile: the position is intended for a highly-motivated applicant who holds a master degree in telecommunication engineering, or applied mathematics or physics. Written communication skills in English are a prerequisite. Applying candidates will prove and/or justify the requested knowledge and skills by providing: ● A full curriculum vitae; ● A motivation letter consistent with the proposed PhD project; ● All documents attesting the requested skills and knowledge; ● Full academic records and marks; ● One or two recommendation letters are considered a plus. Required Skills: The candidate's curriculum must demonstrate solid theoretical competences in modeling and performance evaluation. Knowledge of control theory, Markov theory and machine learning (especially Reinforcement learning, Neural Networks and Federated Learning) is preferred. Experience on event driven simulations is considered a plus.
English: Candidate profile: the position is intended for a highly-motivated applicant who holds a master degree in telecommunication engineering, or applied mathematics or physics. Written communication skills in English are a prerequisite. Applying candidates will prove and/or justify the requested knowledge and skills by providing: ● A full curriculum vitae; ● A motivation letter consistent with the proposed PhD project; ● All documents attesting the requested skills and knowledge; ● Full academic records and marks; ● One or two recommendation letters are considered a plus. Required Skills: The candidate's curriculum must demonstrate solid theoretical competences in modeling and performance evaluation. Knowledge of control theory, Markov theory and machine learning (especially Reinforcement learning, Neural Networks and Federated Learning) is preferred. Experience on event driven simulations is considered a plus.
2025-09-15
Apply
Close
Vous avez déjà un compte ?
Nouvel utilisateur ?
More information about ABG?
Get ABG’s monthly newsletters including news, job offers, grants & fellowships and a selection of relevant events…
Discover our members
MabDesign
Institut Sup'biotech de Paris
SUEZ
Aérocentre, Pôle d'excellence régional
TotalEnergies
MabDesign
ADEME
Groupe AFNOR - Association française de normalisation
ONERA - The French Aerospace Lab
Généthon
Nokia Bell Labs France
CESI
ASNR - Autorité de sûreté nucléaire et de radioprotection - Siège
CASDEN
PhDOOC
ANRT
Ifremer
Tecknowmetrix
Laboratoire National de Métrologie et d'Essais - LNE
-
JobRef. 132696Montreal, CanadaMcGill University
Post-doctoral position in medicinal chemistry
Scientific expertises :Chemistry - Biochemistry
Experience level :Junior
-
JobRef. 132742Genève, SwitzerlandEPSU
Professeur de Biologie UP à Genève
Scientific expertises :Biology - Biochemistry - Chemistry
Experience level :Confirmed